MRM P&C - Automated Model Risk Scoring

Purpose

MRM P&C - Automated Model Risk Scoring is designed to quantify model complexity, usage, and exposure in order to assign model risk scores for insurance-related model risk management activities. It also uses GenAI to explain the rationale behind each rating.

Primary users

The primary user is not specified in the provided information. The agent is associated with the AQS team and is intended for users involved in model risk scoring or review where model complexity, usage, exposure, and rating rationale need to be assessed.

Where it fits (process/stage/trigger)

This agent fits into a model risk management process for P&C insurance models, particularly when model metadata, performance data, and risk usage logs are available for assessment. The specific process stage or trigger is not specified in the provided information.

Key capabilities / workflow

The agent analyzes model complexity, usage, and exposure, then assigns a risk score based on the provided model-related data. It also generates an explanation of the rationale for each rating using GenAI, supporting transparency in the scoring workflow.

Inputs

Typical inputs include model metadata, performance data, and risk usage logs. Other input requirements are not specified in the provided information.

Outputs / Deliverables

Outputs include assigned model risk scores and GenAI-generated explanations of the rationale for each rating. Other output formats or deliverables are not specified in the provided information.

Value

The agent supports more consistent and explainable model risk scoring by combining quantitative assessment of model complexity, usage, and exposure with generated rating rationale. This can help model risk stakeholders better understand why a model receives a given risk rating.